Genetic material is successfully introduced usually only into a very limited number of target cells of a population. This necessitates the distinction and isolation of successfully transformed from nontransformed cells, a process which is referred to as selection. Traditionally, the selection is carried out by way of a “positive” selection, wherein the transformed cell is enabled to grow and to survive, whereas the untransformed cell is inhibited in its growth or destroyed (McCormick et al. (1986) Plant Cell Reports 5:81-84). A positive selection of this kind is usually implemented by genes which code for a resistance to a biocide (e.g. a herbicide such as phosphinothricin, glyphosate or bromoxynil, a metabolism inhibitor such as 2-deoxyglucose 6-phosphate (WO 98/45456) or an antibiotic such as tetracycline, ampicillin, kanamycin, G 418, neomycin, bleomycin or hygromycin). Such genes are also referred to as positive selection markers. The positive selection marker is coupled (physically or by means of cotransformation) to the nucleic acid sequence to be introduced into the cell genome and is then introduced into the cell. Subsequently, the cells are cultured on a medium under the appropriate selection pressure (for example in the presence of an appropriate antibiotic or herbicide), whereby the transformed cells, owing to the required resistance to said selection pressure, have a growth/survival advantage and can thus be selected. Positive selection markers which may be mentioned by way of example are:                phosphinothricin acetyltransferases (PAT) (also: Bialophos® resistance; bar) acetylate the free amino group of the glutamine synthase inhibitor phosphinothricin (PPT) and thus achieve a detoxification (de Block et al. (1987) EMBO J. 6:2513-2518; Vickers J E et al. (1996) Plant Mol Biol Reporter 14:363-368; Thompson C J et al. (1987) EMBO J 6:2519-2523).        5-enolpyruvylshikimate 3-phosphate synthases (EPSPS) impart a resistance to the unselective herbicide Glyphosat® (N-(phosphonomethyl)glycine; Steinrucken H C et al. (1980) Biochem Biophys Res Commun 94:1207-1212; Levin J G and Sprinson D B (1964) J Biol Chem 239:1142-1150; Cole D J (1985) Mode of action of glyphosate; A literature analysis, p. 48-74. In: Grossbard E and Atkinson D (eds.) The herbicide glyphosate. Buttersworths, Boston.). Glyphosate-tolerant EPSPS variants for use as selection markers have been described (Padgette S R et al. (1996). New weed control opportunities: development of soybeans with a Roundup Ready™ gene. In: Herbicide Resistant Crops (Duke S O, ed.), pp. 53-84. CRC Press, Boca Raton, Fla.; Saroha M K and Malik V S (1998) J Plant Biochemistry and Biotechnology 7:65-72; Padgette S R et al. (1995) Crop Science 35(5):1451-1461; U.S. Pat. No. 5,510,471; U.S. Pat. No. 5,776,760; U.S. Pat. No. 5,864,425; U.S. Pat. No. 5,633,435; U.S. Pat. No. 5,627,061; U.S. Pat. No. 5,463,175; EP-A 0 218 571).        neomycin phosphotransferases constantly impart a resistance to aminoglycoside antibiotics such as neomycin, G418, hygromycin, paromomycin or kanamycin by reducing the inhibiting action thereof by means of a phosphorylation reaction (Beck et al. (1982) Gene 19:327-336).        2-deoxyglucose 6-phosphate phosphatases impart a resistance to 2-deoxyglucose (EP-A 0 807 836; Randez-Gil et al. (1995) Yeast 11:1233-1240; Sanz et al. (1994) Yeast 10:1195-1202).        acetolactate synthases impart a resistance to imidazolinone/sulfonylurea herbicides (e.g. imazzamox, imazapyr, imazaquin, imazethapyr, amidosulforon, azimsulfuron, chlorimuron ethyl, chlorsulfuron; Sathasivan K et al. (1990) Nucleic Acids Res 18(8):2188).        
In addition, resistance genes to the antibiotics hygromycin (hygromycin phosphotransferases), chloramphenicol (chloramphenicol acetyltransferase), tetracycline, streptomycin, zeocine and ampicillin (β-lactamase gene; Datta N, Richmond M H. (1966) Biochem J 98(1):204-9) have been described.
Genes such as isopentenyl transferase (ipt) from Agrobacterium tumefaciens (strain:PO22) (GenBank Acc. No.: AB025109) may likewise be used as selection markers. The ipt gene is a key enzyme of cytokine biosynthesis. Its overexpression facilitates the regeneration of plants (e.g. selection on cytokine-free medium) (Ebinuma H et al. (2000) Proc Natl Acad Sci USA 94:2117-2121; Ebinuma H et al. (2000) Selection of Marker-free transgenic plants using the oncogenes (ipt, rol A, B, C) of Agrobacterium as selectable markers, In Molecular Biology of Woody Plants. Kluwer Academic Publishers). The disadvantages here are, firstly, the fact that the selection disadvantage is based on usually subtle differences in cell proliferation and, secondly, the fact that the plant acquires unwanted properties (gall tumor formation) due to transformation with an oncogene.
EP-A 0 601 092 describes various other positive selection markers. Examples which may be mentioned are: β-glucuronidase (in connection with, for example, cytokinine glucuronide), mannose 6-phosphate isomerase (in connection with mannose), UDP-galactose 4-epimerase (in connection with galactose, for example).
Negative selection markers are used for selecting organisms in which marker sequences have been successfully deleted (Koprek T et al. (1999) Plant J 19(6):719-726). In the presence of a negative selection marker, the corresponding cell is destroyed or experiences a growth disadvantage. Negative selection involves, for example, the negative selection marker introduced into the plant converting a compound which otherwise has no action disadvantageous to the plant into a compound with a disadvantageous (i.e. toxic) action. Examples of negative selection markers include: thymidine kinase (TK), for example of Herpes simplex virus (Wigler et al. (1977) Cell 11:223), cellular adenine phosphoribosyl transferase (APRT) (Wigler et al. (1979) Proc Natl Acad Sci USA 76:1373), hypoxanthine phosphoribosyl transferase (HPRT) (Jolly et al. (1983) Proc Natl Acad Sci USA 80:477), diphtheria toxin A fragment (DT-A), the bacterial xanthine-guanine phosphoribosyl transferase (gpt; Besnard et al. (1987) Mol. Cell. Biol. 7:4139; Mzoz and Moolten (1993) Human Gene Therapy 4:589-595), the codA gene product coding for a cytosine deaminase (Gleave A P et al. (1999) Plant Mol. Biol. 40(2):223-35; Perera R J et al. (1993) Plant Mol Biol 23(4): 793-799; Stougaard J; (1993) Plant J 3:755-761; EP-A1 595 873), the cytochrome P450 gene (Koprek et al. (1999) Plant J 16:719-726), genes coding for a haloalkane dehalogenase (Naested H (1999) Plant J 18:571-576), the iaaH gene (Sundaresan V et al. (1995) Genes & Development 9:1797-1810) or the tms2 gene (Fedoroff N V & Smith D L (1993) Plant J 3: 273-289). The negative selection markers are usually employed in combination with “prodrugs” or “pro-toxins”, compounds which are converted into toxins by the activity of the selection marker.
5-Methylthioribose (MTR) kinase is an enzyme whose enzymic activity in plants, bacteria and protozoa, but not in mammals, has been described. The enzyme may convert an MTR analog (5-(triromethyl)thioribose) as a “subversive substrate” of the methionine salvage pathway via an unstable intermediate to give the toxic compound carbothionyl difluoride.
Said selection systems have various disadvantages. The introduced selection marker (e.g. resistance to antibiotics) is justified only during transformation and selection but is later a usually unnecessary and often also undesired protein product. This may be disadvantageous for reasons of consumer acceptance and/or approval as a food and/or feed product. Another disadvantage in this connection is the fact that the selection marker used for selection is usually genetically coupled to the nucleic acid sequence to be inserted into the genome and cannot be decoupled by segregation during propagation or crossing. Usually, deletion of the marker sequence is required, making additional steps necessary. In addition, biotechnological studies require in numerous cases multiple transformation with various gene constructs. Here, each transformation step requires a new selection marker unless the previously used marker is to be laboriously deleted first. This, however, necessitates a broad palette of well-functioning selection markers which are not available for most plant organisms.